OpenAI says one of its models produced a solution to the Navier-Stokes equation. The equation describes how fluids such as water and air move. It is one of the Clay Millennium prize problems, each worth $1 million. Sebastien Bubeck of OpenAI said training on a new maths-focused model began on 28 August. After hearing rumours that Anthropic was close, OpenAI put far more computing power on the problem. The proof was formalised in Lean, a language used to check mathematical proofs. Tristan Buckmaster of NYU says OpenAI rushed ahead after learning of his work with Anthropic researcher Levent Alpöge. He also says OpenAI proposed a paper that omitted Alpöge's name. Bubeck and Sam Altman both denied that. OpenAI denies inspecting the pair's Codex prompts, but wrote that it cannot rule out that de-identified usage data improved its models.

• 1,000 agents over 50+ hours
• up to 10,000 agents
• compute cost in the millions of dollars
• $1 million Clay prize